Computational protocol: Fungal community profiles in agricultural soils of a long-term field trial under different tillage, fertilization and crop rotation conditions analyzed by high-throughput ITS-amplicon sequencing

Similar protocols

Protocol publication

[…] After sequencing, raw data were imported into an in-house processing pipeline [] for sequencing primer and barcode trimming. Data analysis was conducted by separation and amplicon barcode trimming of pooled ITS1 and ITS2 datasets. This processing step was performed based on the FASTX toolkit (http://hannonlab.cshl.edu/fastx_toolkit/). Separated datasets were further processed and analyzed by a pipeline as recently described [–]. Briefly, raw sequences were merged by FLASH []. Sequences with > 1 N (ambiguous bases) in the read and expected errors > 0.5 were discarded. The molecular identifier tags and primer sequences were removed allowing 0 and 2 mismatches, respectively. The software package UPARSE was applied for de-noising and chimera detection []. Operational taxonomic units (OTUs) were defined at 97% sequence similarity by applying the program USEARCH 8.1 []. Processed OTUs were taxonomically classified using the RDP classifier 2.12 [] with the UNITE database v7.0 [,]. Only hits featuring a confidence value of at least 0.8 (phylum rank) were considered. Finally, raw sequences were mapped to the OTU sequences to receive quantitative assignments. Identified genera were compared to entries in the “List of Plant Diseases” published by The American Phytopathological Society. The lists of wheat (actual crop) and rapeseed diseases (pre-crop) were selected (www.apsnet.org/publications/commonnames/Pages/default.aspx) and names of the fungal pathogens (anamorph, teleomorph and synonymous names) adjusted with the index fungorum website: http://www.speciesfungorum.org/Names/Names.asp. Maize diseases were neglected since no evidence for fungal infections was reported for the experimental site analyzed. [...] Factorial ANOVA for wheat yield was calculated using SPSS Statistics v22 (IBM, Germany). Least significant differences (LSDs) with p≤0.05 for a strip-split-plot design were estimated according to Thomas []. All OTUs recovered from ITS1 and ITS2 datasets were subjected to principle component analysis (PCA) using ClustVis []. For hierarchical clustering based on “City Block” distance and average linkage, Cluster 3.0 was used []. PCA was conducted to further describe the similarity between the fungal communities of different samples. Differentially abundant OTUs were identified by application of the metagenomeSeq Bioconductor package (v1.1.16), []. The approach implements a novel normalization technique and a statistical model addressing under-sampling in high-throughput sequencing projects. The bias in the calculation of differential abundance introduced by total-sum normalization is corrected by cumulative sum scaling (CSS) []. Secondly, under-sampling of microbial communities is reassessed by implementation of a zero-inflated Gaussian distribution mixture model []. The first step was a non-parametric permutation test on t-statistics. From each replicate, 40,000 sequence reads were randomly picked in 100 permutations followed by calculation of average diversity indices considering all permutations. Secondly, a non-parametric Kruskal-Wallis test was performed followed by the Wilcox rank-sum tests on subgroups to avoid positive discoveries of differentially abundant features driven by potential cofounders. PermANOVA was performed using PAST [] and the Shannon and Simpson indices were calculated based on in-house perl scripts inside the pipeline. Venn diagrams were created with Venny []. The analyses of significant differences in the datasets were calculated with SigmaPlot (Systat Software, San Jose, CA). For selected putative phytopathogenic or plant beneficial fungal genera raw data were transformed into ranks to achieve normalized distribution for Three-Way-ANOVA (Shapiro-Wilk, p = 0.05; Brown-Forsythe, p = 0.05) with post-hoc test Holm-Sidak (p = 0.05). Krona plots (– Files) to zoom into taxonomic assignments of treatment replicates were designed by using the interactive metagenomic visualization web browser []. […]

Pipeline specifications

Software tools FASTX-Toolkit, UPARSE, USEARCH, RDP Classifier, SPSS, ClustVis, metagenomeSeq, VENNY, SigmaPlot, Krona
Applications Miscellaneous, Phylogenetics, Metagenomic sequencing analysis, 16S rRNA-seq analysis
Organisms Fungi, Triticum aestivum, Zea mays
Diseases Pulmonary Fibrosis